convection simulation

[super gnu]

ive downloaded the demo and im trying to do a 2d simulation of airflow caused by interaction with hot and cold objects.

i did attempt this before a couple of years ago but the demo used to be severely limited in cell count so i could never get a decent result. i have now conveniently forgotten how i did it.

any pointers to get me started? im testing an idea for convective cooling of a house using a solar chimney.

is phoenix suitable to basically test these kinda things? im not after precise results, just a general visualisation of the effect.


initial questions:

*how would i set up a scene where air temperature can be visualised in an approximate 10-100 degree temperature range..? i remember i had problems with this before..

*how do i assign a temperature to scene elements

*is phoenix capable of simulating things like the venturi effect?


any pointers most appreciated..

[super gnu]

ok found my old thread..

https://forums.chaosgroup.com/forum/...mal-simulation

however im unable to recreate the simulation i did.. im getting no convection.. see attached file. in reality the hot air rising at the top would suck the lower "cold" air horizontally into the space. in this scene however, the lower cold source behaves identically with or without the hot source active. either im making a nooby mistake or youve "fixed" your undesirable thermal diffusion.

[georgi.zhekov]

Hello,

It's the velocity that should affect the fluid, not the temperature.
Also if you introduce some cooling in your scene it will create some negative pressure.
Adding more conservation quality would help your scene as well - try with 50 or 100.
For the rendering - don't forget to set your Smoke color and Smoke Opacity to be Based on Temperature - here is how this looks.

[super gnu]

thanks, the conservation quality was the one i was missing, however , while i now get the "suction" effect i wish in this version, when i do another version (see attached scene) i get an odd result. In this new scene the heat cannot escape the space (its closed at the top) , so the hot air should gradually fill the structure from top down, forcing the cooler air out of the lower vent.

however despite starting well ( the lower source begins venting out of the structure) after a number of frames it reverses direction and floods into the space, at the same time as the heated air is filling the space from the top. obviously impossible..!

this is with a conservation quality tested all the way to 2000.

[Svetlin.Nikolov]

Hey,

A few hints might be helpful:
- You should turn off vorticity when you are doing experiments that should not be affected by random velocity fluctuations. You can find out more about it here: https://docs.chaosgroup.com/display/...mics-Vorticity
- You should put the cold source in Brush mode instead of Surface Force - otherwise it would keep pumping velocity in all directions and this would of course disrupt the flow. You can find out more about source modes here: https://docs.chaosgroup.com/display/...mittedChannels
- If you want a strong conservation with less velocity dissipation, you should switch to PCG. Forward transfer is a good choice when you want to get more accurate results.

I also lowered the temperature for the lower source to 0 and used conservation quality 1000 to exaggerate the effect and exclude some random noise from interfering with the results.

vid.zip

Cheers!